I also suspected that shorter chains saponify first, and it is a shame that no studies address this issue...
I have some experience of this in isolating specific mono-unsaturated fats from seed oils.
In short saponification of mixtures triglycerides is not that simple to model. The rate of the reaction is dictated by how intimate the contact between the aqueous hydroxide and the hydrophobic triglyceride is.
Granted the smaller chains will, in principal, be favoured in the reaction but this would not be exclusive in a mixture (i.e. no fractional saponification is possible). The modelling of this process becomes much more complex once the fats begin to saponify, as the first few percent which is converted to "soap" aids the formation of an emulsion (micelles) of the remaining triglycerides and water so the reaction becomes much more chaotic.
If you wanted to fractionate fats then you would be better off making methyl esters (FAME) and fractionally distilling them to separate into low - medium - and high boiling FAME's which may give you the difference in soap properties that you are looking for.